Search

The spatial QRS-T angle (QRS-Ta), a non-invasive risk marker for arrhythmia and sudden cardiac death, captures information on the spatiotemporal dynamics of ventricular depolarization and repolarization. QRS-Ta peak is the angle between QRS and T-wave loop maximal amplitudes. We compared QRS-Ta peak with intra-cardiac unipolar electrocardiogram parameters simultaneously recorded in the right ventricle, left ventricle (LV) endocardium and LV epicardium (coronary sinus) in 10 patients with structurally normal hearts. S1S2 restitution protocols were performed by pacing from the LV at intervals decrementing from 1000 ms to the effective refractory period (ERP). Repolarization time (RT), activation time (AT) and activation-recovery interval (ARI), a standard surrogate for local action potential duration, were calculated using standard definitions. Decreasing cycle length (CL) correlated with an increase in QRS-Ta. Two phases were identified 1) A stable QRS-Ta between CLs 1000 to 400ms and 2) a subsequent rapid increase in QRS-Ta with further decrements and a small decrease in QRS-Ta just prior achieving ERP. When plotted against the pacing interval, the QRS-Ta distribution mirrored the repolarisation restitution curve. The QRS-Tangle inversely correlated with mean RT (-0.59, -0.81 / -0.34, P=0.023) and mean ARI (-0.72, -0.73 / -0.42, P

Genome-wide association studies (GWAS) have discovered hundreds of genetic loci for resting heart rate (RHR). However, the impact of intra-individual variation in RHR on GWAS results is unclear. We evaluated this impact by analyzing two RHR recordings from N ~61,000 subjects from UK Biobank. In addition, we modelled variations in RHR as independent white zero-mean Gaussian noise with a standard deviation of 0.5x, 1x, and 2x the standard deviation of the difference between the original RHR values (4,8, and 16 bpm, respectively). The two original RHR recordings were highly correlated (ρ =0.77), but results from the genetic analyses were s lightly different: the number of genome-wide significant (p < 5x10−8) variants at the locus with the strongest reported association (MYH6): n=39 vs. n=34; the p-value of the corresponding lead-variant, 3.6x10−24 vs. 2.1x10−19; and the estimated heritability 20.0% vs. 16.7%. Simulated data showed an inverse relationship between RHR variation and genetic association strength and heritability. Results formally demonstrate the impact of intra-individual RHR variability on the discovery of genetic variants in single-measurement studies.

Coronary artery disease (CAD) is the main contributor to cardiovascular mortality in developed countries, making accurate diagnosis of utmost importance. We developed risk scores to assess CAD risk in a population without known cardiovascular disease by combining ECG and a genetic risk score (GRS) for CAD. We analysed data in 52,260 individuals in the UK Biobank study. ECG indices included heart rate, PR, QRS, QT and T-peak-to-T-end intervals, while we built the GRS from publicly available genome-wide association results for CAD that were derived in an independent population. In a training set (N = 39,195), the indices with the strongest CAD prognostic impact were the PR and QT intervals, and the GRS. When combined together into a Multivariate model, both the ECG markers and the GRS were independently associated with CAD. In an independent test set (N = 13,065), we then built three risk scores based on (1) ECG markers, (2) genetic data, and (3) a combination of ECG and genetic data, respectively. The hazard ratio (95% confidence interval) for CAD comparing high versus low-risk individuals was 6.5 (5.1 – 8.3),8.4 (6.4 – 10.8) and 8.4 (6.5 – 10.8) for the three risk scores, respectively. In conclusion, the inclusion of genetic markers into risk scores with ECG markers independently contributes to CAD risk prediction in a large population of individuals without known cardiovascular disease.

The unipolar electrogram (UEG) is widely used in electrophysiological research and in the cathlab. We aimed to test a previously proposed simple model of the UEG against in-vivo human data and to use the model to investigate: (A) Accuracy of repolarization measurements; (B) Factors affecting UEG substrate mapping and (C) Interactions between APD and UEG T-wave alternans. UEGs were recorded in 10 patients using a multi-electrode sock. Local action potentials showing same activation and repolarization sequence as measured in-vivo were generated using analytical functions. Local UEGs were simulated as the difference between the local action potential and a position-independent component representing remote activity. Morphological correlation between recorded and simulated UEG was cc = 0.92 (0.79 – 0.97) (median Q1-Q3, N = 1, 756). Simulation studies showed: (A) Caution should be used when analyzing biphasic T-waves and T-waves associated with either very early or late repolarization. (B) Substrate mapping using UEG amplitude depends on the activation sequence and its total duration. (C) UEG TWA is not a specific surrogate for local APD alternans as it can be observed in sites without APD alternans due to variations in the remote component. In conclusion, the simple model provides a framework to improve the understanding and clinical utility of the UEG.

Success rate of ventricular tachycardia (VT) ablation remains sub-optimal. Current technology does not allow fast and accurate delineation of the ablation target. Noninvasive panoramic ECG-imaging (ECGI) offers the possibility of studying the interaction between arrhythmogenic substrate and earliest sites of activation during VT to improve ablation strategies. ECGI mapping (CardioInsight, Medtronic) was performed in 5 patients undergoing VT ablation. Ventricular pacing was delivered from the RV and three indices were measured at each ventricular site to map susceptibility to arrhythmia initiation: Re-entry vulnerability index (RVI), local dispersion of AT (ΔAT) and local dispersion of repolarization (ΔARI). Regions of high susceptibility were defined as those corresponding to the bottom 5% of RV I and the upper 5% of ΔAT and ΔARI. Morphologically distinct VTs were analyzed to measure the AT sequence and localize the region of earliest epicardial activation (AT 0.13 for all pair-wise comparison). The vulnerable region presented at least partial overlap with the region of earliest activation during VT in 50%, 55% and 50% of all VTs for RV I, ΔAT and ΔARI, respectively. These early data confirm the mechanistic link between markers of arrhythmogenic risk and VT initiation and suggest that ECGI could be potentially used for targeting ablation in non-inducible or hemodynamically non-tolerated VTs.

Heart rate (HR) response to exercise and recovery are strong predictors of cardiovascular mortality, but the HR profile morphology may add useful information for the identification of subjects at risk. Our aim was to characterise the HR profile morphology and assess its cardiovascular risk predictive value. 1-lead ECG recordings of 17,691 participants from the general population in an exercise stress test from the UK Biobank study were analyzed. A methodology based on time warping of the HR profile was applied to compute the average HR profile morphology along the exercise test. Then, two series of warping-based morphological differences in amplitude, $d_{a}$ , and time, $d_{w}$ , were calculated by comparing each individual HR profile morphology with respect to this average HR profile. Subjects who suffered one or more cardiovascular events showed significantly lower values of $d_{a}$ than survivors (median of −10.5% vs −7.5%, p=0.009). Also, $d_{a}$ was significantly associated with cardiovascular mortality in a Cox model after adjusting for clinical variables, resting HR, difference between peak and resting HR, or between peak and recovery HR after a follow-up period of five years (p

Dispersion of ventricular repolarization (DRT) is an important factor contributing to the vulnerability to life-threatening arrhythmias. An accurate non-invasive methodology for its estimation would contribute to improve risk-prediction. We assessed 3 multi-lead ECG markers to track changes in DRT using intra-cardiac data recorded in patients with structurally normal ventricles. Changes in DRT were measured with intra-cardiac unipolar electrograms (UEG) simultaneously recorded in the RV endocardium (RVendo), LV endocardium (LVendo) and LV epicardium (coronary sinus, LVepi) in 10 patients. Standard S1S2 restitution protocols were conducted by pacing from the RVendo (n = 8), LVendo (n = 10) and LVepi (n = 7). DRT was measured as latest minus earliest re-polarization time (RT). In the surface ECG, DRT was estimated from precordial and augmented limb leads as: (1) Interval between the earliest and the latest maximum up-slope of the T-wave (ΔT up ); (2) Interval between median T-peak and median T-end (T pe,med ); (3) Interval between the earliest T-peak and latest T-end (T pe,range ). Intra-patient correlation with DRT changes was higher using ΔTup (0.79, 0.66 - 0.89) than T pe,med (0.61, 0.14 - 0.76, $P$ = 0.001) or T pe,med (0.71, 0.44 - 0.79, $P$ = 0.054).

In this study a Heart Rate Variability (HRV) analysis guided by respiration to evaluate different patterns of Autonomic Nervous System (ANS) in response to a cognitive stressor between Major Depressive Disorder (MDD) and control (CT) subjects is presented. Cardiorespiratory Time Frequency Coherence (TFC) reveals the local coupling of HRV and respiration signal which is essential and usually not included in estimation of ANS measures derived by HRV. Parasympathetic activity of ANS is measured as the power at the frequencies where TFC between HRV and respiration is significant, whereas sympathetic dominance is measured as the normalized power in the low frequency band [0.04,0.15] Hz of HRV excluding the power of those frequencies related to respiration. Results showed significantly lower (p < 0.05) sympathetic dominance in MDD with respect to CT subjects during stress, suggesting that ANS reactivity as response to stress stimuli is lower in MDD patients. The study of ANS reactivity to a stressor may serve as a biomarker useful for the early diagnosis and monitoring of MDD patients.

Ventricular tachycardia (VT) recurrence after catheter ablation remains frequent and improved ablation strategies are needed. The re-entry vulnerability index (RVI) is an activation-repolarization marker to localize critical sites for VT initiation. Its use is limited since current electro-anatomical mapping systems (EAMS) cannot provide global measurement of activation and repolarization times within a single beat. We carried out a simulation study to assess a simple method to measure RVI using data collected by sequential EAMS and we investigated the effect of background noise, RT variability (σ RT) and ectopics on RVI estimation. The mean correlation coefficient between single ECG beats and a representative template is used as inclusion/exclusion criterion. Localization of the vulnerable region associated with 5% bottom RVI was accurate (sensitivity 80±8%, specificity> 99± 1%) for moderate to large repolarization variability (5 ≤ σRT ≤ 20 ms) and moderate level of noise (S N R ≥ 10 dB) but it deteriorated for σRT ≥ 25 ms and S N R 0.67 ± 0.05, MAE < 25 ± 1 ms). The number of ectopic beats did not affect the results. In the in-vivo case analyzed, the sites of low RVI and VT exit was close (5.1 mm).

Identification of targets for catheter ablation of arrhythmias remains a significant challenge. Traditional mapping techniques often neglect the tissue repolarization, hampering the detection of pro-arrhythmic regions. We have recently developed a novel mapping procedure, termed the Reentry Vulnerability Index (RVI), which incorporates both activation (AT) and repolarization (RT) times to identify ablation targets. Despite showing promise in a series of experiments, the RVI requires further development to enable its incorporation into a clinical protocol. The goal of this study was to use computer simulations to optimize the RVI procedure for its future usage within the clinic. A 2D sheet model was employed to investigate the behavior of the RVI algorithm under mapping catheter recordings resembling clinical conditions. Conduction block following premature stimulation was induced and mapped in a cardiac tissue model including repolarization heterogeneity. RVI maps were computed based on the difference between RTs and ATs between successive pairs of electrodes within a given search radius. Within 2D sheet models we show that RVI maps computed on irregular sparse recording sites were in good agreement with high resolution maps. We concude that the RVI algorithm performed well under clinically-relevant mapping conditions and may be used to guide ablation.

The re-entry vulnerability index (RVI) is an activation-repolarization marker recently proposed to predict sites of ventricular tachycardia (VT) initiation. RVI is inversely related to the probability of establishing a re-entry. The aim of this study was to characterize the CL dependency of RVI, assess different methods for CL-dependency corrections and test the capability of RVI to predict ventricular arrhythmias. Twenty-four subjects underwent whole heart epicardial mapping using a multi-electrode sock enabling the recording of 240 unipolar electrograms. Ventricular pacing was delivered at CLs decreasing from 600 to 350 ms in steps of 50 ms. In a separate study, 1 patient went into VT during steady state pacing. Predisposition to VTwas assessed by using the 10th percentile RVI, termed global RVI. The results show that own to CL dependency of local repolarization, there was a strong positive association between RVI and CL. Local repolarization detrending and correction with the Bazett's formula eliminated the CL dependency, while a weak association was found after correction with the Fredericia's formula. In the patient who developed VT, global RVI was significantly lower than in the patients who did not develop VT. Corrections for CL dependency enhanced these differences. In conclusion, de-trending and Bazett's corrections effectively compensated for the CL dependency of RVI and global RVI may reveal predisposition to ventricular arrhythmias. Further analysis is necessary to establish the role of RVI for risk stratification.

Accurate cardiac risk stratification is crucial for preventing cardiac death, but still remains an unmet need. Mechanical alternans (MA), an oscillation of blood pressure that occurs on a beat-to-beat basis, has been recognized as a marker of cardiac instability and is associated with an increased risk of cardiac death. However, the use of MA for risk stratification is currently limited by the invasiveness and costs of continuous blood pressure monitoring. A non-invasive, cuffless and affordable method to detect MA is therefore highly desirable. The pulse transit time (PTT) and the pulse arrival time (PAT) are promising techniques for continuous blood pressure monitoring, particularly for detecting short-term blood pressure changes. In this work, we hypothesized that PAT, measured as the interval between the R-wave in the ECG and a distal arterial pulse, can be used to accurately track fast beat-to-beat blood pressure dynamics and detect MA. A total of 42 ECG and femoral arterial pressure recordings from 12 patients with normal ventricles were analyzed. Patients were instructed to breath at a fixed respiratory rate and MA was induced by ventricular pacing. Both MA and PAT alternans were detected using spectral analysis. MA was present in 69% of blood pressure recordings (n=29). ROC analysis showed that PAT accurately detected MA, with an area under the curve was equal to AUC = 0.94. The optimal threshold for detecting MA by using PAT provided 90% sensitivity and 85% specificity. In conclusion, this study demonstrates that PAT can be used to accurately detect pacing-induced MA and may represent a first step toward non-invasive, cuff-less and affordable MA screening for cardiac risk assessment.

The spatio-temporal organization of cardiac repolarization modulates the vulnerability to dangerous ventricular arrhythmias. Methodologies that provide accurate assessment of cardiac repolarization are of primary importance for a better understanding of cardiac electrophysiology and represent a potentially useful tool for clinical applications. The most commonly used repolarization time (RT) marker from extracellular recordings is derived from the unipolar electrogram (UEG). However, far field potentials and remote activity may in certain conditions bias this marker. In this paper, a RT marker based on the bipolar electrogram (BEG) is proposed. An analytical expression of the BEG based on a simple model of the cardiac extracellular potential is derived. According to the proposed analytical framework the BEG exhibits a repolarization wave whose extremum (maximum or minimum) corresponds to the average of the local RTs at the two electrodes of the bipole. The amplitude of this extremum is a function of the steepness of phase 3 of the action potentials, inter-electrode distance, conduction velocity and direction of wave-back propagation. A simulation study based on this analytical framework showed that for noisy to good signal quality (SNR of the UEG ≥ 10 dB), and for a typical inter-electrode distance of 2 mm, conduction velocity between 0.2 and 0.6 m/s, and an angle between conduction direction and the inter-electrode axis ≤ π/4, the median absolute error was lower than 6.8 ms while the median linear correlation between estimated and theoretical RT was higher than 0.91. Examples of RT derived from BEG recorded in a structurally normal heart in both the right and left ventricles demonstrate that the proposed procedure is feasible in human in-vivo studies.

The effect of music on cardiovascular dynamics may be useful in a variety of clinical settings. The aim of this study was to assess whether listening to music characterized by different emotional valence affected ventricular periodic repolarization dynamics (PRD), a recently-proposed non-invasive index of sympathetic ventricular modulation. The 12 lead ECG was recorded in 71 healthy volunteers exposed to six 90 s excerpts of pleasant music and unpleasant acoustic stimuli as well as six 90 s intervals of silence. A 20 s interval was allowed between excerpts during which the participants were asked to evaluate the previous excerpt. A simulation study was carried out to assess the capability of the algorithm of tracking fast small changes in PRD. The simulation study shows that the algorithm implemented in this study has a time-frequency resolution sufficient to capture the fast dynamics observed in this study. PRD were higher during listening to both pleasant and unpleasant music than during silence. There was a (weak) trend for the PRD to be higher during listening to pleasant than unpleasant music that may indicate the existence of a (weak) interaction between the valence of music-induced emotions and sympathetic ventricular response. The PRD significantly increased during the 20 s interval in between conditions, possibly reflecting a sympathetic response to the evaluation task and/or to the expectation of the following excerpt.

The comparison between the pointwise amplitude of different T-waves provides insight into ventricular repolarization liability. However, T-wave pointwise amplitude variability can be confounded by time-domain variability. We, first, compared two algorithms for removing (warping) time-domain variability, one using the original and another one using a transformed T-wave (SRSF). We, next, compared the robustness against noise of two markers, dy and da, of pointwise amplitude variability, after warping the underlying temporal variability with the preferred warping algorithm. dy was obtained from the transformed T-waves while da was proposed in this work and was derived from the original T-waves. We, finally, used the most robust marker to measure the T-wave pointwise amplitude variability between every T-wave recorded during a Tilt test and their mean T-wave. Results showed that the preferred warping algorithm was the SRSF because it is not affected by differences between the amplitudes of the original T-waves. In addition, the marker da presented lower relative error values than dy for every level of noise. The analysis of electrocardiogram records showed that da was significantly lower during the tilt than in supine position (5.5 % vs 6.5 %, p

Precision medicine has been heralded as an opportunity to improve risk prediction, driven significantly by an increasing availability of genetic data. Genetic testing for rare mutations linked with Mendelian monogenic syndromes is available in specialised clinics. For complex diseases however, aggregation of common and low frequency variants into a “polygenic risk score” (PRS) is necessary due to their small individual effect sizes. PRSs for coronary artery disease (CAD), hypertension and atrial fibrillation have shown some modest success at a population level. However, scepticism remains whether the genetic effects in CV disease are sufficient to have meaningful clinical impact. This review explores recent efforts to utilise genomic data for risk prediction using CAD as an example.